May 14, 2007

Contextual In-Video Advertising: ScanScout

Duncan Riley

15 comments »

scanscout.pngVideo advertising is a final frontier in the monetization of Web 2.0. Adbrite was first to market with Adbrite In-Video. Last week Google started testing in-video text ads on YouTube.

Cambridge, MA based ScanScout joins the fray. Founded 2 years ago, ScanScout launched recently and has followed up with an announcement of $7million Series A funding in a round led by General Catalyst Partners.

At first look, ScanScout’s video advertising product looks identical to what Google is testing with YouTube. Text ads are overlaid on the video and open video-on-video advertisements or external sites.

We don’t know a lot about the tech behind Google’s offering. I noted in my post covering the subject last week that the sample YouTube advertisements lacked context. ScanScout on the other hand has no question mark on the issue.

ScanScout technology scans each video and determines content, with ads delivered contextually to match each scene. Think of it as an Adsense for video because it’s exactly how it works, but on scenes as opposed to pages.

I’m yet to be convinced that text based overlays are the future of online video advertising. ScanScout argues that pre-roll and post-roll are regarded as dead by many because they “leverage an old paradigm that essentially ignores the consumer”, and yet this optional form of advertising can easily be ignored itself.

If in-video text advertising is indeed the future of online video advertising, contextual delivery is essential and ScanScout provides a product that delivers exactly that.

scanscout1.png

  • Sphere It

Trackbacks/Pings (Trackback URL)

  1. ScanScout, sistema de publicidad contextual en videos
  2. Blog de software » Blog Archive » ScanScout, sistema de publicidad contextual en videos
  3. Adsense For Video, Google Still Lagging
  4. Kiptronic Adds Advertising Insertion For Video
  5. Web2NewYork » Blog Archive » YouTube is kicking tires
  6. Video Ads: Every Startup Has A Different Solution

Comments

RSS feed for comments on this post.

  1. Steve S

    Very cool. Yea, I think text is still the way to go with online ads.

    Steve,
    Howtosplitanatom.com

  2. doc-film-net

    This is an interesting idea and we have been looking for something like this.

  3. fo.unta.in

    “scanning video” - i think nobody is actually scanning video. the just use the audio in the clip to identify the ads to display. even blinkx does only audio.

    this is the first step. but eventually, some technology will come up to actually scan video. then it will get interesting.

    but seeing what happened to riya (image recognition), this is doubtful.

  4. luciano

    Hi, on my blog I have more link text ads clicked than the videos…funny!

    http://www.lucianobove.blogspot.com

  5. Ali

    I think text-overlay is the wrong way to go as well. I would much prefer a side tab that either pops up “in browser” {with AJAX or something} on rollover. So I control what I want to see.

    Give more control to the user and they’ll love you for it.

  6. sumsum

    do they plan to pay something to the creator of the video content – or is it another “free ride” on content creators ?

  7. fukeo

    @ Fountain.

    I don’t think many people are even “scanning” audio, more like the surrounding text “metadata” and then presenting that to an adsense-like engine?

  8. youtubesearcher

    “ingorability” should not be the key benchmark for determining if and advertising model will work.

    After all, probably the most ignorable ads on the whole internet are Google adwords. And what existed at the time Google adwords came out ? an increasing loud/flashing/blinking bunch of ads, all try to not be ignored.

    No, the key benchmark should be relevance.

    Not sure about ScanScout, but having continuous, subtle, relevant advertising seems right to me.

  9. pallet jack

    - I think scanning the audio is something I could program in a month or so.

    - mash up the “you talk it types” into a txt file …. then feed that into a ad serving context advertising model. Easy

    - Now to have something actually watch the video (besides human) is amazingly hard.